What is zero-click search and why does it matter for B2B content strategy?

Zero-click search is a query that ends on the results page, with the user reading a synthesised answer rather than visiting the source. It reached 68 percent of Google searches in the first four months of 2026, and AI Overviews now appear in 58 percent of queries, up from 12 percent in 2024, cutting click-through rates by close to 60 percent where they show. For a B2B company, this means ranking well no longer guarantees traffic: a page can hold position one and still receive a fraction of the visits it would have delivered in 2023.

If publishing more content no longer drives traffic, what should replace the content calendar?

A calendar built on publishing volume assumed clicks would follow rankings and that production cost would keep competitors' output limited. Neither holds: 95 percent of B2B marketers now use AI in their workflow, yet fewer than four in 10 report improved performance, and 43 percent say they cannot differentiate their content in a saturated market. Third Hemisphere typically replaces volume targets with a narrower set of high-authority assets anchored to named executives and distributed through owned, earned, and platform channels under the Waterfall Strategy, so each asset amplifies the next rather than competing with it for the same search real estate.

How should companies measure content performance now that Google referral traffic is falling?

Referral traffic is no longer a reliable proxy for content performance; Chartbeat recorded a 34 percent fall in Google search referrals across more than 2,500 news sites between December 2024 and December 2025. Third Hemisphere's Proprietary Reputation Monitoring and Intelligence Platform tracks share of voice inside AI-generated answers, sentiment, and executive visibility alongside traditional traffic, giving clients a clearer read on whether content is shaping the narrative rather than simply sitting on a page nobody clicks through to. This shift matters most for executive commentary, where 96 percent of B2B brands publish some form of thought leadership but only 4 to 11 percent rate their programme as advanced.

What does a content strategy engagement with an agency like Third Hemisphere typically cost?

Total fee for a six-month term: AUD 90,000, payable in monthly instalments of AUD 15,000, is a representative starting point for a content strategy engagement covering executive commentary, message architecture, and distribution planning across owned and earned channels. Scope and fee scale with the number of executive spokespeople, markets, and platforms involved, and Third Hemisphere sets this out in a fixed-scope proposal before work begins rather than an open-ended commitment. Fees are structured as a fee for the agreed term rather than an ongoing charge, which keeps the scope and the deliverables tied together.

How long does it take to see results from a content strategy rebuilt for the AI Overview era?

Most engagements run six to twelve months, with a fortnightly WIP, monthly reporting, and a quarterly review built in to track shifts in share of voice and executive visibility rather than short-term traffic spikes. Early signals, such as inclusion in AI-generated answers or pickup by named trade publications, typically appear within the first quarter, while measurable shifts in market positioning and inbound enquiry usually take two to three quarters to consolidate. This timeline reflects the pace at which search engines like Google update their AI Overview sourcing and re-index authority signals, not an arbitrary agency milestone.

How do you choose a partner to rebuild content strategy after zero-click search?

The right partner treats zero-click search as a distribution and measurement problem rather than a reason to publish faster, since 43 percent of B2B marketers already say mass-produced AI content has made differentiation harder, not easier. Agencies like Third Hemisphere in Australia and Singapore combine message architecture, named executive positioning, and a Waterfall Strategy that anchors every channel to the same core assets, so content built once earns visibility across search, social, and trade media rather than being reproduced at volume. Look for a partner who can show how they measure presence inside AI-generated answers, not just clicks, given that Datos and Similarweb-based analyses disagree on the exact zero-click rate but agree on the direction of travel.

Should B2B companies stop investing in content marketing given the fall in Google referral traffic?

No; the fall in referral traffic argues for a change in what gets funded, not a withdrawal from content altogether. Content Marketing Institute data shows AI tools leading 2026 budget increases at 45 percent, while Salesforce found teams running AI-assisted content and sales development workflows cut cost per lead by 38 percent, evidence that efficiency gains are real even as raw output loses value. Third Hemisphere typically advises redirecting budget from volume production toward fewer, higher-authority pieces tied to named executives and distributed through the channels most likely to be cited inside AI Overviews and trade coverage.

Content Strategy After Zero-Click Search Hit 68 Percent

68 Percent of Google Searches Ended Without a Click. Your Content Calendar Is the Wrong Response

Two things happened to content marketing in the same 12 months, and together they finished off the model most companies still fund.

The first is distribution. Zero-click searches reached 68 percent in the first four months of 2026, the fastest acceleration the measure has recorded in a decade. Google AI Overviews now appear in 58 percent of queries, up from 12 percent in 2024, and where an AI Overview appears, click-through rate falls by close to 60 percent. Chartbeat tracked Google search referrals across over 2,500 news sites and recorded a 34 percent fall between December 2024 and December 2025.

The second is production. 95 percent of business-to-business marketers now use AI somewhere in their workflow, and fewer than four in 10 say it has improved performance. The cost of producing a competent 1,200-word article has fallen close to zero, for every company at once.

A content calendar built on publishing volume assumed both of those conditions were stable. Search sent clicks, and production cost enough to keep the field small. Neither holds. The response most marketing teams are making is to publish faster, which addresses the input that stopped being scarce.

What is zero-click search, and how far has it actually gone?

Zero-click search is a search that ends on the results page, with the user reading an answer rather than visiting a source. It has existed since featured snippets arrived. AI-generated answers accelerated it, because a synthesised paragraph resolves more query types than a snippet ever did.

The honest caveat is that measurement disagrees. Similarweb-based analysis puts zero-click at roughly 69 percent of all queries, rising to 80 to 83 percent on queries that trigger an AI Overview. Datos, using stricter clickstream methodology, recorded United States zero-click rates falling from 24.5 percent in December 2025 to 22.4 percent in March 2026. The gap comes from what each dataset counts as a search and as a click.

The direction is consistent even where the level is disputed. SparkToro's analysis concludes that under one third of Google searches now send a click to an external site. Publisher data agrees: small publishers running 1,000 to 10,000 daily page views lost 60 percent of their Google referral traffic across two years, and mid-size publishers lost 47 percent.

For a company blog, the practical reading is that ranking has been decoupled from traffic. A page can hold position one, be summarised inside the AI answer above it, and receive a fraction of the visits the same position delivered in 2023.

Why did content volume stop working in the same year?

Because everyone got the same production capability in the same quarter.

Volume was a defensible strategy while writing was expensive. A company that published four well-researched articles a month outpaced a competitor publishing one, because the gap reflected real investment. Once a competent draft costs a prompt, the gap closes and the field fills. 43 percent of business-to-business marketers now report struggling to differentiate their content in a market saturated by mass-produced AI output.

The commentary category shows the same compression. 96 percent of business-to-business brands publish some form of executive commentary, and only 4 to 11 percent of those programmes rate themselves as advanced. Trust has moved with it: only 4 percent of marketers consider AI-generated content highly trustworthy without human oversight.

Where AI has produced measurable returns, it has done so in workflow rather than output. Salesforce's 2026 research found teams running AI-assisted sales development and content workflows cut cost per lead by 38 percent. The efficiency is real. The volume is worth less than it was.

Spending patterns have yet to catch up. When the Content Marketing Institute asked where budgets were increasing in 2026, AI tools led the list at 45 percent, which funds faster production of the asset class that lost its distribution.

What still holds value?

One asset class has held its value through both shifts: content that contains information which exists nowhere else.

The mechanism is straightforward. A summariser can compress an explainer, because the underlying facts sit in 40 other articles. A summariser cannot manufacture a number that only one organisation holds. Where a claim has a single source, that source gets named, and the naming is the value.

The evidence supports this directly. Research from Princeton and Georgia Tech found that adding original statistics and cited data to content improves AI citation rates by 30 to 40 percent. Muck Rack's Generative Pulse study, which analysed over 25 million links cited across ChatGPT, Claude, and Gemini in 17 industries, found earned media accounts for 84 percent of AI citations while paid and advertorial content accounts for 0.3 percent.

Three formats sit at the top of what earns coverage and citation: original data studies, expert reactive commentary tied to live news, and free tools or calculators. Each shares a property. None can be produced by summarising what already exists.

Where does original research come from without a research budget?

The common objection is that original research means a commissioned survey costing tens of thousands of dollars. Sometimes it does. Often the material already sits inside the business, unrecognised because it was collected for another purpose.

Four sources cover most cases.

Operational data. Aggregated, anonymised patterns from the product, the platform, or the service delivery. Average time to resolution, seasonal demand curves, regional variation, failure rates. Companies dismiss this as internal reporting, and journalists read it as a story about an industry.
Structured customer knowledge. Fifty discovery calls, coded and counted, produce a defensible finding about what buyers in a category are actually struggling with. The rigour comes from counting consistently rather than from sample size alone.
A built benchmark. Assess 100 competitor websites, 50 job advertisements, 200 council planning applications, or every product in a category against one clear criterion. The work is tedious and the output is unique, which is the entire point.
A structured read on a regulatory or policy change. Detailed sector-specific analysis of what a change does to costs, timelines, or obligations, published before consultants have priced it.

Each of these produces something a competitor cannot replicate with a prompt, and each gives a journalist the one thing they need, which is a number nobody else has.

How many assets does a company actually need a year?

Fewer than most calendars assume.

A workable pattern for a mid-sized company is four substantial evidence-led assets a year, one per quarter, each with original data at its centre. Around each, a cluster of derived material: an executive byline arguing what the finding means, a media release for the data itself, a set of social posts, a webinar or briefing, and updates to the pages the finding supports.

That produces roughly 20 to 30 published items a year, and the entire output traces back to four pieces of genuine work rather than 30 separate attempts to have something to say. It also survives the distribution problem, because a data study is pitched to media and cited by AI engines rather than depending on organic search to find it.

The reallocation is usually cost neutral. Most content teams already spend the equivalent of four quarterly studies on producing 50 articles that no longer earn the traffic they were built for.

How do you measure a content programme that no longer chases sessions?

Sessions and rankings stay on the dashboard, and they no longer sit at the top of it. Three measures matter more now.

Citation presence: does the finding get named in AI answers to the questions it addresses, and does the company get named alongside it? Earned pickup: how many independent publications carried the data, and are they publications the target buyer reads? Sales usage: does the sales team send the study, and does it appear in deals?

The third is the least fashionable and the most reliable. A study that the sales team volunteers is doing commercial work regardless of what the analytics say.

How Third Hemisphere approaches content strategy

Third Hemisphere is an Australian communications agency working with founders, startups, and organisations across technology, climate, deep tech, and financial services. Its content strategy work starts by finding the proprietary evidence a client already holds, turning it into something publishable, and taking it to media rather than posting it and hoping search delivers an audience.

That approach follows from how the agency is built. Media relations, content, and AI search visibility sit together, so an asset is designed from the start to earn coverage and citation rather than to fill a calendar slot. Further reading sits in the agency's insights and resources library, and the approach page sets out how the pieces connect. Companies reviewing a content programme for next year can book a free consultation.

The takeaway

Search stopped delivering clicks and AI stopped making content scarce, in the same year, which leaves publishing volume as an expense with no distribution behind it. The content that still works contains information that exists in one place, and every company has more of that than it realises. Four pieces of real evidence a year, taken to market properly, outperform 50 articles competing against an infinite supply of the same thing.